vertical profiles figure

[teixeirak]

@mcgregorian1, I'm envisioning a single plot with about 8 panels (4 cols x 2 rows) showing (a-d) NEON vertical profiles, (e) crown positions, (f-h) trait averages (depending on which traits come out in top model).

add panels for:

• [x] height by crown position in 2018 (order by height, not alphabetical):

  

• [x] trait averages (can have more than one trait per plot)

general:

• [x] across all plots, add a dashed horizontal line indicating the 95th percentile of height.
• [x] make consistent y-axis
• [x] condense by displaying only one copy of legend, remove repeated values on y-axis, etc.

for the NEON part:

• [x] add standard deviations

[mcgregorian1]

I'm not sure I understand.

For height, if you use 20 bins, the top bin would represent the 95th percentile, matching your line. For all of these, let's plot on linear axes, putting the point at the median value for each bin. Let's have lines connecting the points, as you do with the other height profile figures.

This is what the graph looks like with 20 bins of equal number of observations. The top bin does represent the 95th percentile, but I'm unsure why we should add the line going through? Also, why would putting these points on the median matter if the numbers are not whole numbers on the axes?

Also, for TWI, please flip the axes.

Flipping the axes creates a scenario like this:

[teixeirak]

My suggestion is to the median points on a continuous axis rather than treating them as discreet. There's no need to display the range of the bin.

[teixeirak]

The line connecting points, and the overall plot type, would be the same as the NEON height profile.

[teixeirak]

The percent missing values was just for us to assess how representative these values may be (no need to display on final graph). I think that we should remove (or plot in grey) categories for which we're missing >25% of the individuals.

[mcgregorian1]

Is this what you're thinking of? The first graph is showing the percentiles, whereas the second graph is showing the same bins but with them plotted at the upper part of their bin range.

[teixeirak]

That looks good. The second one is the one we'll want to show.

[mcgregorian1]

Hi @teixeirak

These are the traits graphs for the hydraulic traits that were part of the candidate set for the best model (only TLP and PLA). The TWI graphs on the right have the axes in the way you were asking for earlier.

Based on this, I'm wondering how to combine all the graphs together? The traits x height graphs fit in with the NEON graphs and the crown position graph, but the TWI axes don't match. I'm thinking have 2 separate images, with 1 image containing the 4 NEON plots and the 3 height plots below, and the second image only containing the TWI plots.

[teixeirak]

I agree. I think the TWI plots belong in the SI, not the main paper.

[mcgregorian1]

@teixeirak here is the full figure for the vertical profiles. I think this is more in line with what you were thinking of having?

[teixeirak]

That's generally good, but a few points that won't match New Phyt guidelines:

• a lot of the text is too small
• backgrounds should be white
• insert (a), (b), (c) lettering on each panel

guidelines copied below:

Figures

Refer to all diagrams, graphs and photographs as 'Figures'.
Present at approximately twice the size that they will appear. Ensure that, after reduction, they will be compatible with the double-column format of the journal.
Subdivisions of figures should be labelled with lower case, bold letters (e.g. (a), (b)) and referred to in the text in the form (Fig. 1a), (Fig. 1a,b). Avoid including explanatory material in the figure itself – this should be in the legend.
Present diagrams and graphs on a white background, with lines approx. 0.5 mm thick, any shading inserted as lines or dots, and symbols should be approx. 3 mm across. The preferred symbols, in order, are open circles, closed circles, open and closed squares, and open and closed triangles. The same symbol should be used for the same entity in different figures. Axes should be clearly marked with units in brackets after the axis title.
With photographs, include any scale bars on the picture. Where a figure is made up of several photographs, these should be abutted unless this affects the clarity.
Figure legends should be included with the main text of the paper, after the References. Key information describing each figure should be in the first sentence; the legend should be a self-contained, full explanation of the figure, with the species under investigation and all abbreviations defined.
In general, figures should be kept to a minimum consistent with scientific necessity.
We recommend that you consult the Electronic Artwork Guidelines at http://authorservices.wiley.com/bauthor/illustration.asp.

Colour

There are no colour charges and we welcome colour figures, which will be published in full colour. Figures submitted in black-and-white will be published as such in the online article. Please ensure that where colour figures are submitted, that their legends make reference to colour where necessary. To ensure optimal reproduction of your colour figure, please ensure that the file is provided in RGB format.

[mcgregorian1]

That's generally good, but a few points that won't match New Phyt guidelines:

• a lot of the text is too small

Changing the text size (and still having all the text fit) means I can't have 7 plots within one image. I can make things bigger but then I would need to make separate plots.

• insert (a), (b), (c) lettering on each panel

Do you mean for every single panel or just the main groupings (e.g. NEON, the boxplot, and the trait graphs)?

[teixeirak]

Do you mean for every single panel or just the main groupings (e.g. NEON, the boxplot, and the trait graphs)?

every panel.

[teixeirak]

*

Changing the text size (and still having all the text fit) means I can't have 7 plots within one image. I can make things bigger but then I would need to make separate plots.

I'm confident that it's possible. Here are some tips to make it work.

• increase font size of numbers on axes by quite a bit. It's okay if there are fewer numbers
• for canopy position, abbreviate D, C, I, S, and we can explain in the legend.
• for NEON profiles, legend can be overlaid on top of the final panel
• also for NEON profiles, remove "mean" from X-axis label, make sure its clear in legend
• as necessary, abbreviate x-axis variable names (e.g., air temperature = T_air, where air is a subscript)

[mcgregorian1]

@teixeirak you were right, I had just never seen text manipulation like that before. I think this works much better.

[teixeirak]

Yes, much improved! A few more little things:

• [x] remember to change Percent leaf area to Percent loss area or PLA (better).
• [x] you can also also abbreviate TLP (no need for "mean")
• [x] include boxes around the last 3 plots
• [x] make dotted line at 35 m clearer. You could make it color if you'd like (no charge for color figures).
• [x] arrange legend vertically (as was), but inside last box. Remove "month" and "line type" labels.
• [x] make all font the same size (matching largest)
• [x] no bold on axes labels
• [x] better if you can make subscripts on axis labels
• [x] probably better to use parentheses for units
• [x] put letters in parentheses

Sorry to be picky! The journal (regardless of where its ultimately published) will ultimately come back to you requiring figure formatting, and its really easiest to just to make it right from the start.

[mcgregorian1]

• [x] make all font the same size (matching largest)

so you don't want the axis titles to stand out by being larger?

• [x] no bold on axes labels

Currently I don't have any bold command in the code. I believe that's how the coloration is rendering when I save the image and then take a snip of it to put here. It looks darker because it's a slightly larger font size than the axis tick marks.

• [x] better if you can make subscripts on axis labels
• [x] probably better to use parentheses for units

The formatting from New Phytologist says units must be in brackets. How do you mean for the subscripts?

[teixeirak]

so you don't want the axis titles to stand out by being larger?

No. I've found (from experience redoing figures to meet journal guidelines!) that the best way to go is generally to make all the font the same size (within journal guidelines). Especially when you're trying to condense axis titles on a multi-panel figure, keeping the axis labels even smaller means they will often be too small to read. You can play around a bit when you have space, but generally when you're tight on space you want all font the same size (journal min specification).

[teixeirak]

The formatting from New Phytologist says units must be in brackets.

Okay, nevermind! Sorry.

[teixeirak]

* better if you can make subscripts on axis labels

Like this: T_air (T<sub>air</sub>)

[teixeirak]

• [x] please change T_infrared to T_biological.

[mcgregorian1]

Ohh ok, I see.

I've been checking over the guidelines and it says nothing for fonts of figures; the only thing I can find is having font be consistent throughout the paper, with an example of Times New Roman 12-pt. Based on this I'm planning on making the figures be that (and eventually making the text be that as well).

[teixeirak]

Figures do not need to be Times New Roman 12! In fact, I'm not sure that would fit guidelines.

[teixeirak]

If these guidelines don't specify anything, just stick with your current font.

[mcgregorian1]

I guess I'll keep it as it was. All it says is that larger fonts are easier for reading

[mcgregorian1]

I think this works now. Regarding the bold-looking axis labels, it's only because of the relative size of the text itself. The figure letters (a,b,c) I did command to bold.

[teixeirak]

I agree. Thanks!

[teixeirak]

Closing this. There may be some minor cleanup on this figure later, but no need for an entire issue.